Real-Time Monitoring System Based on Wireless Sensor Networks and Remote Sensing Techniques for Landslide-Prone Areas in the Northern Region of Thailand

Many areas in the northern region of Thailand are mountainous and subject to landslides. A landslide causes local damage to a particular area; in addition, it induces nearby prone areas trans-regionally. Therefore, monitoring landslide-prone areas can help in risk assesment and management to efficiently handle a situation. This paper reports collaborative research activities and results of projects endorsed by ASEAN COSTI and e-Asia JRP. It aims to investigate potential landslides’ environmental parameters and conditions using real-time monitoring based on wireless sensors networks (WSN) and remote sensing techniques. The WSN, equipped with a camera, is designed based on low-power technologies (i.e., ZigBee and LoRa), and the performance is evaluated and compared in terms of communication packet losses. Also, remote sensing techniques are used to detect changes in satellite imagery and 3D point clouds. For example, k-means clustering and structural patch decomposition are applied to generate a binary map that locates change between two satellite images of different times. A deep-learning approach based on CNN is used to identify locations of landslides. The density-based spatial clustering and a clutter-removal method are used to detect rockfall events from two 3D point clouds.